Objectives. Nonunion is one of the most troublesome complications to treat in orthopaedics. Former authors believed that atrophic nonunion occurred as a result of lack of mesenchymal stem cells (MSCs). We evaluated the number and viability of MSCs in site of atrophic nonunion compared with those in iliac crest. Methods. We enrolled five patients with neglected atrophic nonunions of long bones confirmed by clinical examinations and plain radiographs into this study. As much as 10 ml bone marrow aspirate was obtained from both the nonunion site and the iliac crest and cultured for three weeks. Cell numbers were counted using a haemocytometer and vitality of the cells was determined by trypan blue staining. The cells were confirmed as MSCs by evaluating their expression marker (CD 105, CD 73, HLA-DR, CD 34, CD 45, CD 14, and CD 19). Cells number and viability were compared between the nonunion and iliac creat sites. Results. After three weeks, numbers of 6.08×10. 6. cells (. sd. 2.07) and 4.98×10. 6. cells (. sd. 1.15) were obtained from the nonunion site and the iliac crest, respectively, with viability of 87.1% (81.7% to 90.8%) and 89.8% (84.7% to 94.5%), respectively. No differences was found between the two sources of MSCs regarding cells number (p = 0.347) and viability (p = 0.175). Conclusions. Our findings showed the existence of MSCs in the site of atrophic nonunion, at a similar number and viability to those isolated from the iliac crest